DFO PhD Dissertation Defense: Beyond the observer's gaze
Department of Fisheries Oceanography
PhD Dissertation Defense
"Beyond the observer's gaze: an integrated approach to detection, estimation, and mitigation of observer and deployment effects in fisheries monitoring"
By: Debra Duarte
Advisor
Steven X. Cadrin (UMass Dartmouth)
Committee Members
Pingguo He (UMass Dartmouth), Gavin Fay (UMass Dartmouth), Geret DePiper (Texas A&M), and Anna Malak Mercer (NOAA)
Thursday April 30, 2026
1:00 PM
SMAST East 102-103
836 S. Rodney French Blvd, New Bedford
and via Zoom
Abstract:
Observers are deployed on commercial fishing trips to collect representative samples of discard rates. However, fishers may change their fishing behavior when an observer is onboard (i.e., “observer effect”) or observer programs may over- or under-sample portions of the fleet (i.e., “deployment effect”). If the extent of these effects is substantial, observer data will not be representative of unobserved trips, potentially biasing the estimation of discards. This sampling bias can impact catch monitoring, stock assessments, and fishery management. The goal of this dissertation was to evaluate how well we can detect these types of effects, understand their impacts on catch and discard estimates, and explore mitigation strategies. The New England multispecies groundfish fishery was used as a test case throughout.
Chapter 1 examined the performance of several published methods for detecting an observer effect using a simulation of observer and deployment effects at varying sampling ratios (i.e., observer coverage) for several sample statistics. The simplest methods (t-test and F-test for difference of means and variances) provided an accurate but imprecise estimate of the observer effect size and only when there were no deployment effects. A generalized linear mixed effects model (GLMM) was also not reliable for detecting small bias but was not confounded by deployment effects and was relatively robust to changing coverage rates. The most complicated tests involved comparing differences in trip characteristics between subsequent trips for observed-unobserved and unobserved-unobserved pairs. These tests were able to detect smaller observer effects and were not confounded by deployment effects but were unreliable at high coverage rates (>60%), producing both high false positive and false negative rates. Sensitivity tests also showed differing detection accuracy as the distribution of the metric of interest changed. No single method was reliable across all conditions, indicating that the choice of method should depend on the specific characteristics of the fishery.
Chapter 2 compared the impact of observer and deployment effects on catch and discard estimates from multiple methods: stratified ratios, generalized additive models, generalized linear models, and random forest models. Several methods were robust to the impact of deployment effects, but the preferred model differed by species, and variability between iterations was high for some species. When an observer effect reduced only the proportion of catch discarded, models for estimating total catch were relatively unaffected, but discard estimates were underestimated in all models. In contrast, when the observer effect altered fishing behavior (e.g., fishing location or gear configuration), model estimates were biased for both catch and discards.
Chapter 3 created a framework for determining observer coverage needs to meet precision targets for science and management. This framework was used to evaluate tradeoffs between observer coverage and integration of reference fleets with high fidelity data and fewer incentives to change behavior on observed trips, such as electronic monitoring or cooperative research study fleets. The design of the program with respect to observer coverage (equal or unequal for reference fleet participants vs. non-participants) and discard estimation (stratified or unstratified) was critical for accurate estimates, even in the absence of observer effects. A cohesive program must consider tradeoffs of data precision, logistics, quality, cost, and safety. These findings underscore the importance of representative sampling, appropriate estimation models, and thoughtful design to produce accurate estimates for science and management. Observer and deployment effects may be an inescapable outcome of deploying observers on a subset of fishing vessels, but there are viable options for dealing with them. Detection, estimation, and mitigation must be considered together rather than in isolation to avoid biased estimates, which could lead to inaccurate assessments and errors in stock management.
Join Meeting
https://umassd.zoom.us/j/95408579777
Note: Meeting ID and passcode required. Email contact to obtain
For additional information, please contact Callie Rumbut at c.rumbut@umassd.edu
SMAST East 102-103
: 836 S. Rodney French Boulevard, New Bedford MA 02744
Callie Rumbut
c.rumbut@umassd.edu
https://umassd.zoom.us/j/95408579777